Press release-Enabling European leadership in Artificial Intelligence with better rules on Text and Data Mining Boosting AI policy through Copyright
DIGITALEUROPE fully supports the ambitions of the EU Commission to stimulate development and boost investment into Artificial Intelligence (hereafter: AI).
The inclusion of AI as a Key Enabling Technology in the new ‘Re-Finding Industry’ report was warmly received by stakeholders and the European Commissioners. Indeed, as highlighted by Vice-President Katainen in his speech at the Industry Day: “[w]e need to fully embrace digitisation and Artificial Intelligence. We need a true digital single market and a thriving data economy.”
AI, machine learning and data analytics offer vast potential for economic growth, increased competitiveness of Europe on the global scene, and be enablers for a fairer and more resilient society. Their benefits cannot be underestimated, from healthcare diagnostics, supply chain management and sustainable energy management.
At our DIGITALEUROPE Masters of Digital event on 20 February, AI and data analytics were recurring elements of the discussions, highlighted by speakers from the Commission, healthcare sector and the finalists of our SME award. If AI is to be the engine for growth in Europe, then data is the necessary oil.
AI can therefore only thrive if there is the data policy framework to accompany it. In this context, we welcome the Free Flow of Data Regulation and are following the development of the upcoming AI and Data Economy policy proposals expected in the coming months.
However, as pointed out by various stakeholders across SMEs, NGOs and the academic and research community, an important oversight to the overall data and AI strategy remains: the Copyright in the Digital Single Market Directive risks blocking the development of AI through a restrictive and counter-intuitive view on text and data mining (hereafter: TDM).
In particular, article 3 in the proposed Directive would bring most TDM within the scope of copyright legislation and only exempt certain research users. This approach means that everyone other than this group would need to 1) identify the rightholders (thousands with large data sets or even innumerable with web-crawling), 2) request additional approval (even if the content is already available online) and 3) pay an extra licensee fee and possibly have to use the rightholder-provided TDM software rather than their own flexible and fit-for-purpose tools).
This approach ranges from being burdensome and costly, to unsustainable and insurmountable for the incredible range of TDM users, including SMEs, universities, European tech companies and industry. It is also not justified: TDM does not lead to copyrighted content being made available for free, there is no new competing publication or harm to the original creators.
Instead, the proposed article 3 in the Copyright in the DSM Directive would frustrate and complicate TDM, and consequently, the broader data economy and AI development in Europe. The quality of AI algorithms depend on access to large and diverse data-sets. This is recognised for instance in the USA and Japan, where TDM activities are broadly allowed and even encouraged.
It is not too late to act: article 3 of the Copyright in the DSM Directive can be amended to have a broad scope, allowing research and development also for commercial use. This helps European businesses and also academia involved in public-private partnerships. This is also proposed and supported by several Member States in theCouncil’s IPR Working Party, to at least give individual countries this option.
DIGITALEUROPE therefore calls on the Commission, the Member State representatives, and the Members of the European Parliament Members, to turn the positive ambitions towards AI and the data economy into real action, by supporting a broad exception for TDM within the Copyright in the DSM Directive. A thriving and high-quality data market will result in high-quality AI and a competitive European research and development community.